Saved in:
Bibliographic Details
Main Authors: Cai, Kaiwen, Duan, Zhekai, Liu, Gaowen, Fleming, Charles, Lu, Chris Xiaoxuan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.04908
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912053686435840
author Cai, Kaiwen
Duan, Zhekai
Liu, Gaowen
Fleming, Charles
Lu, Chris Xiaoxuan
author_facet Cai, Kaiwen
Duan, Zhekai
Liu, Gaowen
Fleming, Charles
Lu, Chris Xiaoxuan
contents Recent advancements in Vision-Language (VL) models have sparked interest in their deployment on edge devices, yet challenges in handling diverse visual modalities, manual annotation, and computational constraints remain. We introduce EdgeVL, a novel framework that bridges this gap by seamlessly integrating dual-modality knowledge distillation and quantization-aware contrastive learning. This approach enables the adaptation of large VL models, like CLIP, for efficient use with both RGB and non-RGB images on resource-limited devices without the need for manual annotations. EdgeVL not only transfers visual language alignment capabilities to compact models but also maintains feature quality post-quantization, significantly enhancing open-vocabulary classification performance across various visual modalities. Our work represents the first systematic effort to adapt large VL models for edge deployment, showcasing up to 15.4% accuracy improvements on multiple datasets and up to 93-fold reduction in model size.
format Preprint
id arxiv_https___arxiv_org_abs_2403_04908
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Self-Adapting Large Visual-Language Models to Edge Devices across Visual Modalities
Cai, Kaiwen
Duan, Zhekai
Liu, Gaowen
Fleming, Charles
Lu, Chris Xiaoxuan
Computer Vision and Pattern Recognition
Recent advancements in Vision-Language (VL) models have sparked interest in their deployment on edge devices, yet challenges in handling diverse visual modalities, manual annotation, and computational constraints remain. We introduce EdgeVL, a novel framework that bridges this gap by seamlessly integrating dual-modality knowledge distillation and quantization-aware contrastive learning. This approach enables the adaptation of large VL models, like CLIP, for efficient use with both RGB and non-RGB images on resource-limited devices without the need for manual annotations. EdgeVL not only transfers visual language alignment capabilities to compact models but also maintains feature quality post-quantization, significantly enhancing open-vocabulary classification performance across various visual modalities. Our work represents the first systematic effort to adapt large VL models for edge deployment, showcasing up to 15.4% accuracy improvements on multiple datasets and up to 93-fold reduction in model size.
title Self-Adapting Large Visual-Language Models to Edge Devices across Visual Modalities
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.04908